Комментарии:
Excellent tutorial/review. Explained very well with great examples. Thank You👍
ОтветитьAs always, great recommendation! Thank you. I use Topaz DeNoise AI and Topaz Sharpen AI and like the results in 80% of cases.
ОтветитьMatt do you do noise reduction as a first or last step?
ОтветитьThis is a great feature but keep in mind, as someone pointed out to me, the file size increases 3-4 times, so a 30 megabyte file will end up being over 100 megabyte.
ОтветитьIs this a beta version, because I don't see it on my LrC CC? Or is this a third party plug-in?
ОтветитьThank you Matt.
ОтветитьThese Nextorage cards are:
Unknown at Amazon in Europe
Currently unavailable at Amazon USA
QUESTIONS:
Sustained read speed - unknown
Burst write speed - different speeds depending on card model and capacity
Sustained write speed - seems to be perfect as demonstrated in Matt's video, but precise number unknown
Capacity
TBW - total bytes written - each write of bits to a memory cell means that that cell is worn. How many times can that be repeated?
Temperature operating range - unknown
IOPS - unknown
Data error protection - unknown
Safe storage time without power - How long can data remain safe in the card when off power? Quantum physicists: between 7 hours and 7,000 hours. SO, WHAT IS IT?
NOTES:
As to TBW. Manufacturers/vendors are mystical about that. It indicates what quality memory cells is being used.
As to speed - it is possible to get higher speed while causing reduced TBW if the wear increases at higher speeds.
As to capacity. Manufacturers/vendors have been very conservative in the past and used "overprovisioning" as a way to up the TBW ante. While the card's controller applies "wear levelling" by storing data in a round-robin way, we can never store more usable data on it than indicated on the label, but having 50% more space under the hood means we get a lot more TBW. As manufacturers/vendors do not get a lot of warranty claims for loss of data, they have started to remove storage capacity out of "over provisioning" and added that to what's available in the label, while at the same time raising the price per GB of storage for exactly the same product now with a different label and "mor capacity".
Yes, naive consumers are being helixed [1] in the market of these products.
[1] a helix is a screw
Well, if you ask the software to guess at detail lost in the noise and smooth out everything in between it does so, no surprise. Just as you can crank saturation up to 11 but normally shouldn't it's up to you to limit this tool to the envelope where it does a good job.
Personally I've used it on a few bird photos taken at ISO 6400 on m4/3 and I'll rather keep it's strength setting under 40. I'd rather have the result showing some fine homogeneous grain than sharp wrinkles in plastic wrap ...
Sorry Matt but this video is better advertised as "Shibari 101" than denoise... I don't like Adobe's denoise compared to other programs but shots in 500 ISO compared with 52000 ISO it is clearly ridicules (as you mentioned that in 6400 ISO was not bad)! Who but a paparazzi will use more than (stretched) 12000 ISO? And most of all for portraits? Disappointing mate!
ОтветитьWhen you see in the thumbnail that the denoised picture is not the same picture… 😂
ОтветитьIt'll be nice when we get a one-click full workflow enhanced by AI. Obviously we'll have to give it an idea what we're looking for with instructions or a preset, but no reason AI can't take over the time-consuming and sometimes tedious job of editing. Probably will happen sooner rather than later.
ОтветитьHave been using DXO Pure RAW 2 on my bird shots for a while now and I can honestly say I'm amazed by the results it produces.
Very straight forward, easy to use interface, can apply lens corrections, sharpen and amazing de-noising, and you can output to a DNG allowing you to still do RAW adjustments afterwards.
With the noise produced by crop bodies like the R7 this software is becoming a necessity to get noise free wildlife images if you're stuck using lower end glass.
I actually don't mind noise. When printed on a 4x6 you can't even see it. Excellent video!
ОтветитьI think considering this is "free" (as in part of the subscription) it's not bad. Of course there are better plug-ins out there but those will cost extra (to some, it's worth it, to others maybe not). But also people do need to realize (as mentioned) it's just a step in the workflow, and I found that sometimes one tool works, sometimes another tool works, and this is why you sometimes have different tools in your photo editing toolbox. Some people can get away with just Lightroom, whereas others need more like Photoshop and possibly even some plug-ins depending on their workflow. For me, I found the LIghtroom Denoise takes too long because my computer is a bit on the older side (4 years old with no dedicated graphics) so the DeNoise AI does take quite a bit of time (usually between 5-15 minutes for most images) but I have other programs I've had that can do it slightly faster so I'll stick with those, but if you don't want to buy the extra plug-ins from Topaz and DXO, etc, this is a good option that you are already getting as part of your subscription. I think while we (most of us) bashed Adobe over the years about adding and improving Lightroom, I think this is yet another example of where some of that profit is going. Over the past 2-3 years Adobe has done a good job at making Lightroom Classic more modernized with things like the advanced masking tools, color adjustment wheels, and now some AI noise removal. Hopefully the "smart" object removal tool in Photoshop also makes it to the release version of Photoshop (I believe it's still in beta).
ОтветитьThat is the best hands-on review of the new AI Denoise filter in Light Room Classic (& ACR) I have seen. Everyone else just ran comparisons at default settings & compared directly with other noise reduction Apps/Plugins currently on the market. No one dared to move a slider, use different subjects at different ISO levels, never mind masking & editing before Saving/Exporting the edited image. What an excellent review Matt! 10/10
ОтветитьWhat do you think about this compared to dxo’s pureraw? It also outputs as dng. There’s a trial, a comparison would be cool!
ОтветитьNot good for the moment, but we can expect it improving in the future
ОтветитьIsn't there a way to mask and then apply denoise? Maybe in a next version. Cause one may want different amounts of denoise in subject v background. I appreciate one can develop a workaround of export and layers in photoshop, etc. just wondering in this version of Lr it is possible in some way.
ОтветитьBeen using it for a few days now, not sure the word "AI" is really suitable, since the denoise seems to be based on some improved algorithms, and I doubt that the algorithm would examine the image and create information originally unavailable by surveying similar images on the internet.
Ответить❤ great system will look into this
Ответить